21 June 2024 in IEEE Transactions on Instrumentation and Measurement
With the widespread use of lithium-ion batteries (LIBs) in electric vehicles, it is essential for battery management systems (BMS) to accurately estimate the state of health (SOH) of the battery. This study constructs a regression estimation model applicable to the SOH of LIBs by combining multi-feature signal analysis with a data-driven approach. First, the collected data were smoothed using advanced filtering methods to extract input features characterising capacity degradation properties from differential thermal voltammetry (DTV) curves, incremental capacity analysis (ICA) curves, meantime decay (MTD) curves, and terminal thermal analysis (TTA) curves. Then, the extracted features are subjected to feature fusion using principal component analysis (PCA) to obtain strongly correlated features with a cumulative contribution of 98% or more as inputs to the data-driven model and the hyperparameters of the group method of data handling (GMDH) network are optimised using simulated annealing genetic algorithm (SAGA). Finally, model construction, validation, and comparison are carried out using NASA and Oxford battery datasets. The RMSE between the actual measured value and the estimated value is obtained to be less than 0.3%, and the R 2 value is as high as 0.99. The results demonstrate that the approach can efficiently catch the phenomenon of capacity regeneration. It is valid and robust for SOH estimation for different types of LIBs.
9 June 2025 in IEEE Transactions on Instrumentation and Measurement
The complex degradation kinetics and inconsistency of decay paths in in-service lithium-ion batteries (LIBs) significantly affect the health management capabilities of battery management systems (BMS). To enhance the performance of state-of-health (SOH) estimation, we propose a Physics-Informed Neural Network (PINN) learning framework. A partial differential equation (PDE) is first formulated based on prior knowledge to model battery cell degradation kinetics. When prior knowledge is insufficient, characteristic variables of degradation are extracted from monitoring data to expand the kinetic model into multiple dimensions. To avoid the direct computation of PDEs for multidimensional models, the Deep Hidden Physical Modeling (DeepHPM) technique is used to construct a function approximator within the PINN framework. Furthermore, a loss function weight-balancing strategy is introduced during model training. By minimizing the negative log-likelihood estimation, the weights are adaptively updated to ensure optimal task balance. The proposed framework is validated on a public dataset and compared against typical machine learning methods through comprehensive analysis. Experimental results demonstrate that the proposed method achieves higher accuracy and reliability in SOH estimation for practical applications, outperforming existing approaches.